NCEP Modeling System Synergy
Meeting Highlights June 27, 2005
The meeting was led by Peter Manousos and attended by
Geoff Dimego, Stephen Lord, John Ward, Mary Hart, Bill Bua, Keith Brill, Joe
Sienkiewicz, and by phone David Bright from SPC and Stephen Jascourt from NWSH.
1. CCS
NCO’s John
Ward reported all systems are up and running – although the development side’s
up time has not been meeting specifications.
The upgrade on BLUE in early June took longer than anticipated. Therefore the upgrade on WHITE is on hold
until the reason for the delay on BLUE can be determined. The RUC 13 upgrade was implemented
operationally as well as a minor change to the GFS. NOTE – all reference to MRF disappears 12Z June 29. (Editor’s
note - The development system had extensive additional down time after the
meeting and experienced loss of critical files, severely curtailing EMC’s model
development and testing).
2. NOTES
FROM EMC
2a.
Global Modeling Group: Stephen Lord reported a new parallel (very long and
involved) including Gridpoint Statistical Interpolation (GSI) will begin in a
week or two. The parallel runs will be
used internally to iron out problems. Output should not be used until official
notice from EMC (around Labor Day).
2b.
Mesoscale Modeling Group:
Geoff Dimego reported the Hi-Res Window upgrade briefing was conducted June 24th and was implemented operationally June 28th. Grid spacing is reduced to between 5 and 6 km and the runs are made with no convective parameterization. Output has not changed (content, resolution, and domain), but output resolution is intended to increase to 5km (NDFD) when more disk space becomes available.
The first phase of generating official Analysis of Record (AOR) is establishing the Real Time Meso Analysis (RTMA). RTMA will use RUC-13 as the first-guess input into a 2DVar version of GSI with anisotropic background error covariances stretched along terrain contours to help downscale to 5-km terrain. RTMA will initially provide a quick analysis for 2 m temperature and dewpoint fields and a 10 m wind field. RTMA data is planned to get into AWIPS in OB7 and may get into the National Digital Guidance Database (NDGD) sooner. Ultimately, when RTMA is fully established it could be run retrospectively to generate the official AOR. Estimates of analysis error will be generated for RTMA fields based on observation density, observation quality, and quality of the first guess field.
The Parallel WRF cycle for the NAM is being run using GSI (since the GSI in its current state is already as good as the Eta 3DVar). This is running at 12km resolution over a domain slightly smaller than the NAM. As with the parallel GFS, the parallel WRF output is not ready for evaluation and users should wait until a formal announcement from EMC (by the end of the summer) to evaluate the output.
2c. Global Ensemble Prediction System: Stephen Lord reported implementation is proceeding for the 6-hrly breeding, vortex relocation algorithm (control of perturbation size and location), and extension of T126 resolution out to 16 days.
2d.
Short Range Ensemble Prediction System: Geoff Dimego reported the addition
of 6 WRF members is intended for the end of September (4th Quarter)
– 2 control runs using the same configuration as in the Hi Res Window
except they will be hydrostatic with 32-km grid spacing and 2 breeding cycles
around these controls, also at 32km resolution. Earlier in September, a minor
bundle is intended for implementation. This will include faster RSM code, a WRF
post unifying RSM and Eta members, allowing BUFR from RSM members and even
ensemble BUFR products. Additionally, output will be extended to 87 hours for
both 09Z and 21Z runs and HI and AK outputs will be created. The forecast
extension and new outputs depend on whether the supercomputer disk space and
scheduling can accommodate it, which is presently anticipated but not
completely certain. Additionally, experiments with using SREF and MREF output
(mean, 10%, and 90% thresholds for some sensible weather elements) to populate
the NDGD are imminent.
2d. Marine Modeling and Analysis Branch (MMAB): No Report.
3. Input to EMC and NCO
OPC - has viewed a
sample of the HYCOMM output and are excited about what they have seen so far.
4. The next meeting will
be held Monday July 25, 2005 at noon in room 209 with remote conference
capability.